Types of Adjudication
There are a number of different ways in which to divide up the many types of adjudication. This article will consider three different ways to categorize or describe an adjudication method. First, adjudication can vary widely based on the amount of input the referee staff has into the process, versus the input from the defined game system. Second, adjudication can vary based on how much a stochastic element can influence the results. Third, adjudication varies widely with the level of focus of a particular wargame. These three different ways of dividing up the various methods and types all begin from the perspective of looking at adjudication of force-on-force warfare (whether it is ground, air, sea, or some combination). Similar divisions could apply to adjudication when it is applicable to other domains within a wargame, based on the principles described here (Table 1).
Table 1- Adjudication Types, and the division they belong to
|Adjudication Type||Dimension Applied To|
|Rigid Adjudication||Formal vs Informal|
|Semi-Rigid Adjudication||Formal vs Informal|
|Free Adjudication||Formal vs Informal|
|Open Adjudication||Formal vs Informal|
|Entity Level||Level of Resolution|
|Aggregated Level||Level of Resolution|
The first look at the taxonomy of adjudication is a way to differentiate per the formality of the system (formal vs. informal). That is, between adjudication where the game system is the final word from other cases, where the referee staff has more of a capacity for subjective input into the process. The names for these types come from Wiggins .
Rigid Adjudication: An adjudication method where the game system (whether it is a manual process or a computer process) is the final word, can be termed a rigid adjudication method. Many hobby wargames (specifically board games and computer games) employ this method, which allows them to be played without a referee. In such a case, the results of the game system are applied, in all situations. These types of adjudication are (if the system works well) fast, because they work without requiring the analysis and input of the referee staff, but only if they can get input reasonably quickly (given the real-world time constraints of hosting a wargame), and produce output reasonably quickly.
Free Adjudication: The other end of the spectrum from rigid adjudication is that of free adjudication. Here, the referee staff observes the decisions and executed actions of the players, and through analysis and subject matter expertise, are able to determine results, and describe them to the players (in essence, creating the next situation in the wargame, for the players to react to). The strength of free adjudication is that it can cover situations that a game system does not predict (and so, cannot adjudicate), and that the results of the referee staff are much more meaningful, in terms of describing success or failure, than an attrition report. The weakness of free attrition, compared to rigid attrition, is that it is first a very much time and labor intensive proposition for the referee staff, and second, very much open to opinion and bias on the part of the referees.
Semi-Rigid Adjudication: In the spectrum between fully rigid games, and free games, there is the idea of a semi-rigid adjudication. This attempts to combine the strengths of both previous types, by allowing a formally defined game system to be used by a referee staff, in order to generate unbiased data-driven results, but still allow for the strength that results from having an informed and flexible referee staff that can adapt to situations outside the scope of the game system. The weakness here is that the time constraints of free adjudication still apply to the referee staff, and they introduce even more delays because now the staff must also be concerned about data inputs, and output generation from the system.
Open Adjudication: Open Adjudication is a method for determining the outcome of conflict through a conversation approach, where the participants are able to describe and defend their own actions, and talk through, as a group, the relative strengths and weaknesses of the competing methods. While this might be very useful in certain situations, in order to have the participants discuss and investigate the potentials within the different proposed and executed actions, it takes on the time management weakness of free adjudication that applies to the referee staff, and exacerbates it by applying the same weakness to the entire set of participants. A variation of this is a Matrix game. The matrix game is a concept invented by Chris Engle, and has all conflict adjudication done by the participants constructing verbal arguments why their actions should succeed. The opposition then produces verbal reasons why the arguments are invalid. Once this is done, the referee assigns a probability of success, and after a dice toss, the results are announced. These types of game have been done at the US Army War College and elsewhere , and are a way to systematize the Open Adjudication method of relying on discussion and argument to adjudicate actions.
The second axis in the taxonomy of adjudication methods presented here lies in the degree of stochastic methods that are presence in the method. This can range from almost no randomness in the case of a deterministic system, to a situation where there is heavy dependence on stochastic influence to the processes employed in the adjudication method. These may be applied to either the rigid or semi-rigid adjudication methods described above. They do not apply to a loose adjudication method, since no system is relied on in such a case.
Deterministic: In the case of a deterministic technique, what is typically done is that there is some a priori evaluation of the likely events to occur within the game design, and for each, a most likely result is described. These deterministic results are then relied on during adjudication. McHugh refers to a deterministic system as an “expected value” system , and it captures the concept of anticipation of the event occurring, and the a priori assigning of an outcome to that event. An example of this sort of technique could be seen when looking at gunfire tables, for instance. If a naval gun is capable of firing 100 rounds in a certain period of time, and it is determined that a .25 chance of each round might strike the target, then damage could be adjudicated based on 25 of the rounds hitting. There are some more complex permutations of this idea, based on a variety of different situations and the application of different methods from statistics and operations research, but the results are generated without resorting to any sort of random number generation.
Stochastic: The opposite of deterministic is non-deterministic, such as a system where instead of having the outcome of a particular situation being predefined (deterministic), it allows for the introduction of a stochastic factor (a random range of possible results). This is a random factor that is introduced for many good reasons. When introduced, it is usually applied to a likely range of outcomes, which may be part of a formula, or could be in a lookup table. The reasons for the random-ness that a stochastic value introduces can be simplified to the fact that no model, regardless of how complicated or forward thinking, can account for all variables that may exist in actual operations. The introduction of the random factor accounts for the fact that the decision maker (player) equally cannot account for all variableness in operations.
The third axis in the taxonomy of adjudication that is given here is the difference resulting from the level of focus, or resolution, of the wargame. The many possible different applications of the term wargame, and wargame system, could apply to different levels of focus, even when we are considering the adjudication of combat actions. It would be possible, in a tactical decision game, for instance, to focus on small units and adjudicate at the level of individual soldiers and vehicles (referred to as entity level, in the LVC simulation community), or to focus on large formations of troops, at the brigade level, or even higher, for a regional or global theater of operations. While the already mentioned differences of the first two axes of the taxonomy of adjudication apply here (they may both be rigid, or free; they may both be deterministic, or stochastic), the interpretation of each of those other differences is also affected by the level of focus. Very coarsely, this axis will look at only two differences, entity level and aggregated level.
Entity Level: Many simulators in the LVC world have sought to introduce greater fidelity into training and analysis by representing combat effects, and adjudicating the results of combat actions, at lower and lower levels (in terms of unit aggregation), which results in higher resolution. In fact, the two factors are typically at odds with each other – great aggregation of units means (necessarily) more abstraction, and lower resolution in the presentation of detail of combat effects and operations. It is beyond the scope of this article to describe the many differences, strengths and weaknesses between low resolution combat models, and high resolution combat models – but both exist, and within appropriate bounds, both could be (and are) used for adjudication, depending on the focus of the game in question. Typically, without resorting to a physics based model that might serve a high resolution first person shooter type computer game, the lowest aggregation is down to the individual entity, or individual combat platform. Adjudication of combat effects at this point revolve around determining the situation of the entities involved in a combat engagement, and then determining the results of that engagement at the single entity level. Typically, this involves some game system that evaluates each entity’s chances of scoring a hit, and then evaluating the results of that hit.
Adjudication at this level of resolution might be useful, especially, for a tactical decision game, but may prove to be too expensive, in terms of compute time and data requirements, for wargames of discovery – unless they are of operations at a very low level. Details on the methods involved, however, are well covered in Strickland  and earlier in Youngren . Methods presented there are very well suited for computerized methods of adjudication, because of the number of non-trivial calculations that have to be performed for even a small engagement. It is worth noting that the non-professional domain of wargaming is rife with very good systems for adjudicating small scale, or skirmish, engagements at an entity level that result in plausible results, very useful for discovery wargaming in a tabletop environment when small units are involved in an engagement. The mechanisms are still related to individual determination of chances to hit, and the effects of a hit, to determine the combat results at an entity by entity rate, but they are typically modeled in such a way that they are able to be performed at a reasonable human pace, rather than at a digital, or computer pace. The history of combat modeling, once it took on a life of its own as a pursuit for the non-professional, led (for instance) to the invention and explosion of table top roleplaying games, which feature a wide variety of detailed rules for determining many aspects of encounters between small groups of individuals, vehicles, and weapon systems .
Aggregated: The complement to entity level combat modeling for adjudication, although the difference resides along a gradual spectrum, is aggregated level combat modeling. This is adjudication of combat actions, so that a determination can be made as to the value of an operation, at an aggregated level of combatants. Typically, this might correspond to military organizational units (battalions, brigades, task forces), but there are also models that take into account the aggregated strength of all units and forces within a single operation, or line of operations, within a campaign. It is possible, using some of the methods of adjudication described below, to work out a campaign based on the entire strength of one side’s military vs the other.
What is lost in aggregated methods, is that the higher the aggregation (i.e. – the larger the group of combat operatives you consider in your evaluation of military operations), the more you have to abstract out details. With reasonably small formations, such as companies or battalions, what gets lost is the idea of the individual. It is not known what each platform or soldier is doing, but that is the point of aggregated combat modeling – you don’t have to know. The abstraction takes all those factors into account, and then the results of opposed combat actions are generated by the game system. This might involve a computerized method, or a manual tabletop method. It might involve calculations involving a stochastic element, or might be based on expected values only, and be deterministic. But the individual action, and to some extent, individual level results, are abstracted out. As the levels of aggregation get larger and larger (for instance, at the level of a brigade, corps, joint task force, or higher) even more detail gets abstracted away. In many respects, this is ideal for wargames of discovery, as the abstracted details may not be needed for the evaluation of courses of action, or determining best case (or novel) responses to particular strategic options. What is important, is to understand the results of the combat action, and the costs (in terms of time, results, and unanticipated consequences).
The means by which aggregated combat is evaluated is done in several ways. By far, for the computerized wargame, one of the more popular methods is the Lanchester Equation, first devised by Lanchester  for studying the effects of air warfare during the First World War, but also ably reported on, and described in depth by Taylor . This is a mathematical algorithm that compares the two bodies of combatants involved in an operation, and by applying certain factors, can determine the levels of attrition that each suffer and inflict, over a series of time steps. This (in many ways) is ideal for a wargame, as it presents the cost of operations (in terms of attrition to each side) over time, giving the players a chance to respond and introduce new decisions and actions. The shortcomings of Lanchester are chiefly two. First is that it involves a series of mathematical formulations that, unless computerized, is extremely time consuming, and may slow down the adjudication staff to an unforgivable pace. The second is that the factors mentioned are extremely difficult to get right, and may have many situational variations, which are difficult to predict and prepare ahead of time.
Lanchester; however, is not the only answer to aggregated combat modeling for adjudication. Two other mathematical methods are worth mentioning, that have been developed in response to Lanchester, and they are Epstein  and Dupuy . Epstein is very much an attempt by computer modeler Joshua Epstein to address shortcomings in applying Lanchester to extremely mobile warfare (such as a situation with extremely efficient methods of movement and mobilization, as would occur with modern nation states, during the Cold War, when he wrote his book). It is an attempt at introducing fixes to Lanchester, but in doing so increments along with many improvements/changes until it is actually a different system. Dupuy introduces a system whereby different types and qualities of units have a different point value, rather than being based on manpower, such as Lanchester, and then introduces methods for determining attrition and effectiveness based on those point values. Again, it is aggregated combat, and the mathematics involved benefit greatly from computerization, but it allows for rapid assembly of the factors involved, by having data dictionaries with lookup tables for the typical units and unit types that might be encountered in a campaign. Such flexibility is very useful for discovery wargames.
Finally, it should be mentioned that in the non-professional world of wargaming, or hobby wargaming, that there are many different methods for adjudicating aggregated combat using techniques that are quite suitable for tabletop and seminar wargames. Classic methods, such as the dice driven combat results table, have been around for many decades, and in some forms, go back to the original data driven combat tables from the Von Reisswitz Kriegsspiel . In the case of the original tables for the Kriegsspiel (in several permutations), this was not dice driven, but fell into the category of a deterministic method, using expected values for attrition over a time period (at that time, for instance, the number of casualties resulting from musketry at a certain range, and over a certain period of engagement). The more typical modern version, such as those originally devised by, and promoted by Charles Roberts for the Avalon Hill Game Company. In that form, the combat results table takes into account the difference between two forces, expressed (usually) as a ratio of force, and then a dice roll introduces variation in results (attrition, retreat, disruption of command, etc). More modern examples include many variations and additional introduced factors that reflect a wide variety of different operational engagement possibilities. The strength of such methods for tabletop or seminar wargames is immediately apparent – they can be executed with relative ease, and in games of discovery where ad hoc reconfiguration of an encounter may be needed to explore alternatives on the fly, such methods are easy to recalculate and reapply. With a more detailed, and more nuanced computerized model the results are much finer in detail, and may produce much deeper results other than simply attrition and disruption, but at the price of not being as flexible, and of course, requiring that the digital equipment be supported (including operation, data support, etc.).